Summary

  • Kiribati’s 1998 telecommunications profile listed satellite links and planned only to investigate Internet service. A later academic article quoted a local entrepreneur who said Telecom Services Kiribati began service at the end of 1999.
  • That respondent described a 128 kbps link to Australia, 32 dial-up modems, more than 400 users and charges of A$32 a month plus A$8 an hour. The same account reported broken modems and repeated dialing, but it is not an independent service audit.

A country can have an international communications route without having a public Internet service that people can readily use. Kiribati’s late-1990s record makes that distinction unusually concrete: a satellite gateway and data channels formed part of the national telecom network, while the first reported commercial Internet offer depended on a separate pool of telephone modems.

The International Telecommunication Union’s 1998 Pacific Islands profile described Telecom Services Kiribati Ltd (spelled “TKSL” in the document) as the national and international service provider. It said international access ran through an IDR satellite link to Sydney and through Telstra’s PACT DAMA network to other Pacific islands. The IDR bearer had sixteen channels, two dedicated to data. Kiribati used eight modems on PACT for access to Pacific neighbours. These were valuable communications facilities, but the report did not describe them as a public Internet subscription service.

Its development plan still listed an investigation into introducing Internet services.

The same 1998 profile shows how much “national connectivity” could conceal. Seventeen outer islands still used high-frequency radio, grouped into six substations, with service on a manual schedule. That is evidence of a geographically uneven telephone network at that date, not a count of Internet users two years later. It also cautions against treating the later Internet connection as if it automatically reached every island or replaced the older links.

The next clear account appears in Trevor H. B. Sofield’s 2002 study, “Outside the Net: Kiribati and the Knowledge Economy.” In a postscript, Sofield quotes an email dated 10 February 2001 from a local IT entrepreneur. The respondent said Telecom Services Kiribati eventually started Internet service at the end of 1999. He described a server bank of 32 modems and a leased 128 kbps backbone to Australia, with more than 400 users. He gave the charges as A$32 per month and A$8 per hour.

Those numbers describe different layers. The 128 kbps figure was the reported capacity of the leased international backbone; it was not a per-subscriber speed. The 32 modems described the local dial-up access bank, a separate limit on how many telephone sessions could be served at once. More than 400 users therefore did not mean 400 simultaneous connections, and it does not tell us how many subscribers connected on a typical day. Neither the operator nor the researcher supplied traffic records, a measured busy-hour rate or a modem utilization series.

A separate ITU table for 2001 estimated one ISP, 2,000 Internet users, 1,000 subscribers and 0.13 Mbit/s of international Internet bandwidth for Kiribati, with no broadband subscribers. The bandwidth estimate is close in scale to 128 kbps, but the table does not explain why its estimate of users differs from the entrepreneur’s “over 400.” The figures may use different definitions or reference dates; the sources do not resolve that gap.

A separate ITU table for 2001 estimated one ISP, 2,000 Internet users, 1,000 subscribers and 0.13 Mbit/s of international Internet bandwidth for Kiribati, with no broadband subscribers. The bandwidth estimate is close in scale to 128 kbps, but the table does not explain why its estimate of users differs from the entrepreneur’s “over 400.” The figures may use different definitions or reference dates; the sources do not resolve that gap.

The respondent also said a couple of the modems were broken and that a user might need to call about ten times before connecting. This is an important service-quality clue, but it remains one person’s reported experience. The article offers no call-completion statistics, repair log or independent test. A damaged port could reduce the number of usable lines below the nominal bank size; repeated redialing could then turn a seemingly available service into a much less predictable one. The sources do not establish how long faults lasted or why they occurred.

One detail must be kept separate. Sofield writes that he himself failed eighteen times over six days to make contact while trying to reach an interview source. That is the researcher’s difficulty making contact. It is not a subscriber survey and it is not the respondent’s estimate that customers sometimes needed about ten calls. Combining the two would make an anecdote look like a measured national failure rate.

The account therefore supports a bounded conclusion. By early 2001, a local respondent reported a commercial service with an Australian backbone, a small dial-up modem bank, several hundred users and usage-based charges. The report also records an access complaint. It does not establish the service’s exact launch date, nationwide coverage, affordability across household incomes, or the quality experienced by all customers. The contrast is still revealing: international capacity could be stated in kilobits, while practical access depended on a working modem, an available telephone line and the cost of remaining online.

The lens in Heng Lu’s Note 64 is useful here only as an editorial question: what is the smallest working interface that participants actually adopt? Applied cautiously, it directs attention from a backbone headline to the operational route between a subscriber and a functioning session. It does not explain Kiribati’s history, and the available sources do not show whether the service was designed around that principle.

Sources